English
Related papers

Related papers: Transcoders Trace Visual Grounding and Hallucinati…

200 papers

A key goal in mechanistic interpretability is circuit analysis: finding sparse subgraphs of models corresponding to specific behaviors or capabilities. However, MLP sublayers make fine-grained circuit analysis on transformer-based language…

Machine Learning · Computer Science 2024-11-08 Jacob Dunefsky , Philippe Chlenski , Neel Nanda

Vision Language Models (VLMs) have achieved impressive progress in multimodal reasoning; yet, they remain vulnerable to hallucinations, where outputs are not grounded in visual evidence. In this paper, we investigate a previously overlooked…

Computer Vision and Pattern Recognition · Computer Science 2025-10-15 Sifan Li , Hongkai Chen , Yujun Cai , Qingwen Ye , Liyang Chen , Junsong Yuan , Yiwei Wang

Medical Vision-Language Models (VLMs) often hallucinate by generating responses based on language priors rather than visual evidence, posing risks in clinical applications. We propose Visual Grounding Score Guided Decoding (VGS-Decoding), a…

Computer Vision and Pattern Recognition · Computer Science 2026-03-24 Govinda Kolli , Adinath Madhavrao Dukre , Behzad Bozorgtabar , Dwarikanath Mahapatra , Imran Razzak

The use of Vision-Language Models (VLMs) in automated driving applications is becoming increasingly common, with the aim of leveraging their reasoning and generalisation capabilities to handle long tail scenarios. However, these models…

Computer Vision and Pattern Recognition · Computer Science 2026-03-09 Nikos Theodoridis , Reenu Mohandas , Ganesh Sistu , Anthony Scanlan , Ciarán Eising , Tim Brophy

Sparse autoencoders (SAEs) improve interpretability in multimodal models, but it remains unclear whether SAE features form modular, composable units for reasoning-an assumption underlying many intervention-based steering methods. We test…

Artificial Intelligence · Computer Science 2026-03-27 Yunpeng Zhou

Vision-Language Models (VLMs) are powerful tools for processing and understanding text and images. We study the processing of visual tokens in the language model component of LLaVA, a prominent VLM. Our approach focuses on analyzing the…

Computer Vision and Pattern Recognition · Computer Science 2025-04-29 Clement Neo , Luke Ong , Philip Torr , Mor Geva , David Krueger , Fazl Barez

Hallucinations remain a persistent challenge for vision-language models (VLMs), which often describe nonexistent objects or fabricate facts. Existing detection methods typically operate after text generation, making intervention both costly…

Computer Vision and Pattern Recognition · Computer Science 2026-03-06 Sai Akhil Kogilathota , Sripadha Vallabha E G , Luzhe Sun , Jiawei Zhou

Understanding the internal activations of Vision Transformers (ViTs) is critical for building interpretable and trustworthy models. While Sparse Autoencoders (SAEs) have been used to extract human-interpretable features, they operate on…

Computer Vision and Pattern Recognition · Computer Science 2026-04-16 Gerasimos Chatzoudis , Konstantinos D. Polyzos , Zhuowei Li , Difei Gu , Gemma E. Moran , Hao Wang , Dimitris N. Metaxas

Vision-language models (VLMs) are powerful but remain opaque black boxes. We introduce the first framework for transparent circuit tracing in VLMs to systematically analyze multimodal reasoning. By utilizing transcoders, attribution graphs,…

Computer Vision and Pattern Recognition · Computer Science 2026-02-25 Jingcheng Yang , Tianhu Xiong , Shengyi Qian , Klara Nahrstedt , Mingyuan Wu

Vision-Language Models (VLMs) increasingly power high-stakes applications, from medical imaging to autonomous systems, yet they routinely hallucinate, confidently describing content not present in the input. We investigate the root causes…

Computer Vision and Pattern Recognition · Computer Science 2026-05-18 Harshvardhan Saini , Samyak Jha , Yiming Tang , Dianbo Liu

Vision-Language Models (VLMs) leverage aligned visual encoders to transform images into visual tokens, allowing them to be processed similarly to text by the backbone large language model (LLM). This unified input paradigm enables VLMs to…

Computer Vision and Pattern Recognition · Computer Science 2025-03-18 Bangzheng Li , Fei Wang , Wenxuan Zhou , Nan Xu , Ben Zhou , Sheng Zhang , Hoifung Poon , Muhao Chen

Object hallucination in Large Vision-Language Models (LVLMs) significantly impedes their real-world applicability. As the primary component for accurately interpreting visual information, the choice of visual encoder is pivotal. We…

Computer Vision and Pattern Recognition · Computer Science 2025-09-18 Weihang Wang , Xinhao Li , Ziyue Wang , Yan Pang , Jielei Zhang , Peiyi Li , Qiang Zhang , Longwen Gao

Many vision-language models (VLMs) that prove very effective at a range of multimodal task, build on CLIP-based vision encoders, which are known to have various limitations. We investigate the hypothesis that the strong language backbone in…

Computer Vision and Pattern Recognition · Computer Science 2025-09-22 Sho Takishita , Jay Gala , Abdelrahman Mohamed , Kentaro Inui , Yova Kementchedjhieva

Large vision-language models (LVLMs) often hallucinate content that is fluent yet unsupported by the image, limiting their reliability in real-world deployment. We show that a key failure mode arises from route competition: even when visual…

Computer Vision and Pattern Recognition · Computer Science 2026-05-26 Zhe Cheng , Wenyu Chen , Fode Zhang , Dehuan Shen

Current visual grounding models are either based on a Multimodal Large Language Model (MLLM) that performs auto-regressive decoding, which is slow and risks hallucinations, or on re-aligning an LLM with vision features to learn new special…

Computer Vision and Pattern Recognition · Computer Science 2025-12-15 Weitai Kang , Jason Kuen , Mengwei Ren , Zijun Wei , Yan Yan , Kangning Liu

Building state-of-the-art Vision-Language Models (VLMs) with strong captioning capabilities typically necessitates training on billions of high-quality image-text pairs, requiring millions of GPU hours. This paper introduces the…

Computer Vision and Pattern Recognition · Computer Science 2025-07-14 Tiezheng Zhang , Yitong Li , Yu-cheng Chou , Jieneng Chen , Alan Yuille , Chen Wei , Junfei Xiao

Autoregressive vision-language models (VLMs) can handle many tasks within a single model, yet the representations that enable this capability remain opaque. We find that VLMs align conceptually equivalent inputs into a shared task vector,…

Computer Vision and Pattern Recognition · Computer Science 2025-05-08 Grace Luo , Trevor Darrell , Amir Bar

Large Vision-Language Models (VLMs) rely on effective multimodal alignment between pre-trained vision encoders and Large Language Models (LLMs) to integrate visual and textual information. This paper presents a comprehensive analysis of…

Computer Vision and Pattern Recognition · Computer Science 2025-11-25 Shweta Mahajan , Hoang Le , Hyojin Park , Farzad Farhadzadeh , Munawar Hayat , Fatih Porikli

Vision-Language Models (VLMs) have demonstrated remarkable performance across a variety of real-world tasks. However, existing VLMs typically process visual information by serializing images, a method that diverges significantly from the…

Computer Vision and Pattern Recognition · Computer Science 2026-02-10 Yueyan Li , Chenggong Zhao , Zeyuan Zang , Caixia Yuan , Xiaojie Wang

Vision-Language Models (VLMs) are increasingly used as perceptual modules for visual content reasoning, including through captioning and DeepFake detection. In this work, we expose a critical vulnerability of VLMs when exposed to subtle,…

Computer Vision and Pattern Recognition · Computer Science 2025-08-14 Jordan Vice , Naveed Akhtar , Yansong Gao , Richard Hartley , Ajmal Mian
‹ Prev 1 2 3 10 Next ›